A Novel Approach for Data Cleaning using CFDs Algorithms

The main objective of this paper is to give idea of conditional functional dependencies. As usage of computers is increasing are day to day, the storage of data also is more. The data has to be removed and cleaned if it is not helpful. Data cleaning should be done repeatedly by finding out the faults. Hence this gave rise to conditional functional dependencies. CFD (Conditional functional dependency) is an integrity constraint. Conditional Functional Dependencies were used to eliminate redundancy. Traditional FDs (functional dependencies) were recently replaced by CFDs for data cleaning. Functional Dependencies were mainly used for schema design whereas CFDs were aimed at capturing the constancy of data by using patterns of semantically or meaningful related constants. Inconsistent Relational data can be identified by using Conditional Functional dependencies. The detection problem is more complicated in the case of CFDs when compared to Functional dependencies also removal patterns in Conditional Functional dependencies will establish more challenges. For discovering CFD two methods are used. First, CFD Miner is used to discover CFDs which have constant patterns. These constant CFDs are important for discovering an object to clean .The second algorithms which were used for determining general CFDs is CTANE, CTANE is considered as expansion of TANE which is used in removal of Functional Dependencies. These algorithms used for cleaning are based on type of application by the user.